Servo Bass

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Some past bass speaker designs have utilised an auxillary winding on the voice coil former and four wire connection to provide servo motional feedback.
Many car subs are available nowadays with twin voice coils.
Does anybody have experience or knowlege of this, or suitable references.
Sony and others have used a cone mounted accelerometer feedback arrangement.
Is there anything suitable and available and economical for retrofitting to existing drivers ?.
And suitable circuits ?.

Thanks and regards, Eric.

BTW - In my archives I have a 1964 National Panasonic all tube Am/Fm reciever with an auxillary coil arrangement and four wire connection, and a front panel switch to control the amount of feedback, or none.
Within the limits of lowish power, okish fidelity amp stages and speakers, it actually works quite well.
IME and IMHO the arrangement on the big Genesis system works pretty well too !. :)
An Acceleration Feedback System from Speaker Builder is available at <a href=""><b>Andrea Ciuffoli Documents section,</b></a> 9 pages in all.

Uses a piezo element from a tweeter I think?

Hope this helps

If you're considering using the second coil for inductive servo feedback, don't forget the effects of the voice coil it's wrapped in/around. The VC will be generating strong magnetic fields that will modulate or even temporarily reverse the permanent magnetic field in the motor. Being a dynamic system, this will be a complex interaction. Unless you can think of a way to filter this, I think you will be more successful with a different or at least separate feedback system.

Feedback systems for woofers score high on my interest meter.
I vote with Bill, interactions between the coils will make it difficult to get a clean signal out of a second voice coil. I've got some ideas that I'll be getting to one of these days, but right now I'm tied up on the electronics end of things.

The adxl150 seem to be a good option, +/- 2v output at 50 or 25g. bandwith 1000Hz. Somevere there is statet that this is just in the value and gravity so that it van bee used vith good result's.
As acceleration is just wat the signal says the speaker should do acc to the signal.

This is in contrast to the induced voltage by a secondary voice-coil, witch also nead's signal treatment, before and in sircuit. Before as an integration to make acceleration bekome speed, the coil speed in a magnet gap produses a voltage ( unloaded ), this is then treated to subtract the induction caused by the driver coil current. I think this is a good point to start with.


maybee the second voicecoil needs to bee loaded to mesure currents all the way?
One type of magnetic MFB sensor that could be built by a skilled hobbyist is the one used on the B&M woofers:

Unfortunately I couldn't find a better picture on the web (although I have better ones on paper !).

The sensor's magnets are glued to the pole-piece. The sensor coil is carried by a small PCB that is glued to the voice-coil former.

The other solution that has already been talked of on this forum is the use of a microphone.


Or glued to a hole in the dustcap, exelent idè! then it is possible to ekstend the throw as the sensor has a longer linear lenght. Anyway i am setting up my system with adxl150 this one has a weight 5g, and would probably be the same as the ekstra coil. And with the coil i need to integrate the signal before my differential amplifikation stage , maybee inkluding dc-servo ahead of the differential stage :( ( including pole rolloff 12/dB okt). The accelerometer gives one less stage for the musik:) = simpler is better :nod: will probably need one servo or other Lf rolloff anyway ..
Or glued to a hole in the dustcap, exelent idè!
I doubt that the dustcap is a suitable place for this.
The dustcap likely isn`t stiff enough and the sensor signal might get falsified by sensing vibrations of the dustcap which are not necessarily related to what the cone actually does.
I believe attaching directly to the voice coil is a far better place .

The adxl150 seem to be a good option, +/- 2v output at 50 or 25g
+/-50g max limit may be is a bit on the too low side. A long throw woofer can easily exceed this limits especially at higher frequencies or high excursion at lower frequencies. See graph below (note: 1g = 9,81m/s*2)

The accelerometer gives one less stage for the musik = simpler is better will probably need one servo or other Lf rolloff anyway ..
I agree on this one. Measuring acceleration directly probably would be the way to go.


  • chart2.gif
    41.9 KB · Views: 1,117

I am new to the diy forum, so everyone else may be able to assume this, but we are talking just about sealed boxes for MF designs? The idea of monitoring and correcting the driver motion in vented (or transmission line) boxes is problematic.

Anway, assuming that we are talking about sealed boxes, I have been building subs using one voice coil of a dual vc woofer as a feedback source for years. It is simple, inexpensive, and generally effective.


This techniques works best at system resonance. It will linearize the signal (less distortion) and "tighten" (reduce overhang or decay times) the bass sound.

If you are bi-amping, using a global feeback amplifier, and if you can current limit your amplifier during modification and testing so it does not self destruct if you do something woring, it can be very easy to implement. If you will allow an increase in the amplifier gain, it would require just the extra wiring and one resistor. If you want to keep the gain the same, it would require adding one resistor and changing the value of another.

The limitations:

Don't try to go too high in frequency. Too much coupling between coils. The coupling does not test as significant in subwoofers operating below 200 Hz.

In impulse testing I have not found any evidence that you can extend the output to frequencies lower than system resonance. Too high of feedback and you begin to increase the decay duration and complexity instead of decreasing it.

So, from the above we get the general rule: Don't try to use too much feedback. Generally about 6db is a good starting point.

The gain to cost and effort equations seems perfect for diy efforts. Worth the try as a starting point. You can always go more complex later as your confidence and skills increase.

If anyone is interested, I can supply graphs.

The adxl250 sensitivity set low (50g instead of 25g option) 2cm p-p taken with eye measurment at 30 Hz gives 5v p-p output. ( must bee near 50g:cool: ) this is 10v feed to MY element, when lying on the table. Element distortion :eek:are VERRY obvius! Easily seen on my scoope. :nod:
You can do it with the 2nd voice coil and a series resistor for current sensing, but an accelerometer is going to be a lot more linear and you will be able to reduce distortion more. It does require more calcs, though.
is an example of a DVC-feedback sub. His sensing coil is different from the driven coil, IIRC.

BTW, the kit at the link above is perhaps cheaper than you could do it yourself - especially if you factor in time ;)
i have had the electronics world article on feedback using 2x10inch DVCs in sealed box using 2nd coil as pickup-
he said he had 2% THD

his assumption that the VC could be used is a tricky one that i wouldnt assume.

i think he was lucky his coil seems to be good for the pickup!

so many people say it cant work that well.

if any one wants the article scans,i can put them on my site
it has circuit schematics etcetc.

i did have them,but redid site and forgot to upload.

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.